Study On Electrical Lifetime Prediction Of Contactor’s Contact Based On Wavelet Analysis

Contactor is one of the most important operating devices in electric drives and it is mainlyused to switch on and off AC/DC circuit for motors and other power loads. Contact is contactor’smain contact element and the lifetime of contactor’s contact directly affects the reliability ofequipments and security operation of industrial production. Therefore, forecasting the electricallifetime of contactor’s contact is very meaningful. This thesis predicts the electrical lifetime ofcontactor’s contact with wavelet neural network. In this method, we can get the contactor’scontact remaining life through the short-term electrical parameter.Firstly, the paper chooses four pairs AgSnO2contact as test material and the electric contacttest system is used for electrical property test of contact. Then, this paper mainly studies waveletdecomposition and wavelet denoising for contactor’s contact and selects proper denoisingthresholds for soft threshold processing. Finally, it constructs the wavelet neural network modelbased on the traditional BP neural network algorithm and takes arc energy, contact resistance andarc time as input variable to predict the remaining life of contactor’s contact by using waveletneural network and the traditional BP neural network.Predicting the electrical lifetime of contactor’s contact with wavelet analysis and neuralnetwork can broaden wavelet analysis and neural network’s applied range in electrical lifetime ofcontactor’s contact and provide new ideas and methods to analyze and predict electrical lifetimefaster, more effectively and more conveniently.